Several years ago, when I first started researching about “Srivijaya” and quickly realized that a place name in Chinese sources (Sanfoqi) that scholars thought indicated that supposed polity on Sumatra was actually a reference to “Kambuja,” I consulted an excellent 2011 dissertation from Berkeley by Ian Nathanial Lowman entitled “The Descendants of Kambu: The Political Imagination of Angkorian Cambodia.”
At the time, I was trying to find out where the name “Kambuja” came from, and whether or not it was in use in the time I was looking at. From Lowman’s dissertation, I could see that the ruling elite of Angkor did use this term, and that it was, therefore, reasonable that this term would have been used by Chinese a thousand years ago to refer to “Kambuja.”
However, as is often the case when conducting research, I didn’t read Lowman’s dissertation from start to end, but instead, skimmed around and looked for information that related to my research.
I just went back to Lowman’s dissertation, and this time I read a chapter entitled “The Myth of Independence from Javā.” In that chapter, Lowman clearly discusses a lot of the historical information that pertains to the research I have been conducting.
“The Myth of Independence from Javā” is a reference to a story that appears in some Angkorian inscriptions, including one that is considered to record the “founding” of Angkor by Jayavarman II.
To quote Lowman, that inscription records that “After Jayavarman II ‘came from Javā,’ Śivakaivalya accompanied the king to Mahendraparvata (Kulen) where, in 802 CE, a man wise in magical science named Hiranyadama performed a special rite which ‘made it impossible for this [country of] Kambujadeśa (Cambodia) to be dependent on Javā, and made it so that there would be only one king who was cakravartin (‘universal ruler’).’” (118)
Lowman examines this story/myth, and in doing so, engages in an extensive discussion of what and where “Java” might have been.
In doing so, Lowman covers a wide range of theories, but, ultimately he cannot resolve the issue. However, from the perspective of my research, I can see that the different ideas and views he examines actually all make sense and fit together.
Lowman begins with the word, “Java,” itself and notes the following:
“An interpretation of the Old Khmer references to ‘Javā’ must first confront the meaning of the modern Khmer term jvā (pronounced chvea) and its modern Thai equivalent, javā (pronounced chawa). There can be little doubt that these ethnonyms have an etymological root in the Old Khmer term. The problem lies in this word’s unstable territorial and political inflections. Modern jvā may be used to mean the island of Java, but in its most common usage it refers to Cambodia’s Malay minority, a community more associated with a profession (trade) and way of life than with a particular homeland or territory. In the past the idea of Javā space may have been territorially ambivalent, perhaps corresponding to the whole peninsular and archipelagic maritime world to Cambodia’s south. Emergent political identities in the 20th century, particularly “Malaysian” and “Indonesian,” have likely obscured this Javā space even further.” (121)
Lowman later also states that “Javā in the records of post-Angkorian Cambodia and Ayutthaya (Siam) has always signified the Malay World.” (125)
However, Lowman then considers an alternative position, namely that the term “Java” might have referred specifically to island Java. Here he states that this is what George Cœdès claimed, and he notes that the time of the Khmer effort to become independent from “Java” corresponded with the flourishing on island Java of a kingdom under a ruling family known as the Śailendra.
As further evidence, Lowman notes that there is a record in Chinese sources showing that “a band of people from She-po (Java) attacked Tongking in 767” and that “We also read in a late 8th century inscription from Phan Rang that a temple to Śiva in this southern Cham polity was burned by ‘armies from Java (javavalasaṃghaiḥ) coming on ships’ in 787 CE.” (122)
As we will see below, there is strong evidence against understanding the Chinese record above as indicating island Java. In particular, the term that was used there, Shepo (yes, that would have been pronounced something like “Java” or “Jaba” in the past), clearly did not indicate island Java (see my working paper for details), but we’ll get to that below.
For now, let’s just keep in our minds the fact that there are records of attacks on the coast of “Vietnam” and “Champa” in the second half of the eighth century by some “Java/Jaba” people.
Lowman than notes further regarding the Śailendra lineage that:
“The reach of Java’s Śailendra kings as far as north the Bandon Bay region of southern Thailand in this period is suggested in the so-called Ligor inscription originally said to be found in Wieng Sa south of Surat Thani but perhaps from Chaiya, which alludes to the “chief of the Śailendra lineage,” named Viṣṇu and carrying the title of Śrī Mahārāja. This unfinished and undated inscription is composed on the back of a stele inscribed with a eulogy of the king of Śrīvijaya and dated 775 CE. The two texts share a fairly similar script, which appears transitional between Cambodia’s 7th century and late 9th century script.
Hence, the inscription of the Śrī Mahārāja is possibly a late 8th or early 9th century text. Since the title Mahārāja and the lineage (vaṃśa) of Śailendra were characteristics of central Javanese royalty in the late 8th century, there was likely some connection between the peninsular Śrī Mahārāja and the kings of the island of Java.
However, this connection may have been due to shared political culture or to a loose marriage alliance with Javanese kings rather than to real Javanese hegemony on the Malay Peninsula.
As Michael Vickery has observed, the polity of the Śailendra kings was located in the interior of Java nearer to the south coast, oriented away from the maritime world that would preoccupy the eastern Javanese polity of Majapahit in the 14th century. If the Śailendra kings did send out invading forces to the mainland in the late 8th century (and we cannot rule out this possibility), they left behind very little evidence of their venture.
The Ligor inscription fails to provide this evidence, as there is no king named Viṣṇu in the epigraphy of the island of Java. We cannot assume that a king who claimed origins in the lineage of Śailendra and who used the title of Mahārāja was necessarily a Javanese king.” (122-123)
So, Lowman notes that we have evidence of a “Śailendra” lineage on island Java and in southern Thailand in the eighth century. However, he questions the idea that this represents some kind of expansion outward from Java (as some scholars had argued).
He continues this skeptical reading in talking about Chola inscriptions in southern India that again mention the name Śailendra in the early eleventh century.
An early 11th century Cola inscription in Sanskrit and Tamil says that a Buddhist temple at Negapatam was completed by a certain Māravijayottuṅgavarman, who was king of Kaṭāha (on the western Malay peninsula) and of Śrīviṣaya (i.e., Śrīvijaya), and who was of the Śailendra lineage (śailendravaṃśa). Apparently, this lineage, though originating in Java, was by the 11th century esteemed throughout the maritime realm and was applied to various local royal lines.” (123)
Hence, while the name Śailendra appears in various contexts, Lowman cautions against seeing direct connections between these places.
Lowman then considers whether the expansion from “Java/Jaba” that the founder of Angkor sought independence from might have come from the supposed “Srivijaya” on the island of Sumatra.
“If in fact there was a moment of ‘Javā’ expansion from the islands to the mainland in the late 8th century, or at least nominal recognition of its sovereignty in mainland regions oriented towards the sea, another candidate for this polity would be Śrīvijaya (see fig. 12). Śrīvijaya’s position as the predominant Malay polity in the region during this period certainly supports the possibility.
This polity was probably based in Palembang in coastal Sumatra during the 7th century, and its king is known to have made claims, mentioned above, in the peninsula as far north as the Bandon Bay region in 775 CE.
As java signified the Malay language and Malay ethnicity in early modern (post-16th century) Cham, the pillaging “armies of Java coming on ships” cited in the 8th century Phan Rang inscription could have been Malays from areas within Śrīvijaya’s greater orbit.
In Angkorian Cambodia Javā may have been the name for the entire collection of polities of Malay ethnicity on the peninsula and on the island of Sumatra, in which case it would stand to reason that 8th century Śrīvijaya would be known as Javā.
Perhaps in that hazy time period Khmer polities were briefly dependent on Malay chiefs who governed the important ports and lanes of commerce in a loose subsidiary relationship with Śrīvijaya.” (123-124)
I find this passage really interesting in light of the research I have done, which I will explain below.
There is quite a bit of speculating going on here, but at the core of this discussion we can see that Lowman is trying to imagine a possible connection between a “Srivijaya” and a Malay “Java,” however he can’t quite determine what that might have been.
Finally, Lowman considers the case for the Malay Peninsula as the location for “Java/Jaba.”
“Claude Jacques believes that the Javā of the inscriptions was relatively close to Cambodia and that it likely represented a Malay polity or confederation of polities in what is today southern Thailand on the peninsula. This area was certainly in striking distance of Angkorian Cambodia and its garrison town of Lopburi in the Menam Basin.
Jacques sees evidence of a Javā within Cambodia’s territorial reach in the 12th century Preah Khan inscription, which states that Jayavarman VII received tribute from two kings of Champa, the king of Yavana (northern Vietnam), and the king of Javā (javendra).
George Cœdès was convinced that this claim to four tributary kingdoms was merely royal hyperbole. In contrast, Jacques suggests that the passage reflects a certain political reality.” (126)
These are many of the extant ideas/theories of what the “Java” mentioned in Angkorian inscriptions might have referred to. Lowman clearly and logically goes through all of these ideas and weighs their strengths and weaknesses.
Ultimately, none of these theories/perspectives persuades him, and he concludes by stating that:
“From the above discussion we can conclude that Cambodia’s Javā was probably both a reality and a myth. While a polity of this name existed from Cambodia’s perspective in the 12th century and likely in the 8th, its history in relation to Cambodian may have been imperfectly remembered and its perceived setting may have changed over time. Nonetheless, a place does not become a myth simply because its existence cannot be definitely confirmed. It becomes a myth when, real or not, it is used to convey intentional meaning, to justify a cause, or to affirm an identity in the present. In this sense Cambodia’s Javā was the most potent kind of myth, as it represented a political other against which Cambodia’s integrity could be imagined.” (130)
Having worked on this topic for years by now, it is entirely clear to me that there was nothing “mythic” about this “Java/Jaba.” There might be elements of the story recorded in Cambodia that were or became mythic, but that’s a different issue.
Instead, Java/Jaba was Cambodia’s main rival throughout much of the Angkorian period, until new rivals – Ayutthaya and Majapahit – arrived on the scene in the second half of the fifteenth century.
Here is what I now know.
In the early centuries AD, there were trans-peninsular trade routes that crossed the northern parts of the Malay Peninsula, such as at the Kra Isthmus and at what is now Nakhon Si Thammarat. At the same time, there are also records in Chinese sources that indicate that some people journeyed around the Malay peninsula, and variations of a term that we can render as some form of “Java” were recorded at that time, although it is not clear where it was, or if these accounts are all about the same place.
At some later point, but certainly by the seventh century, trans-peninsular trade routes across the area of what is now southern Thailand stretching from Phatthalung to Trang, and Songkhla to Kedah in what is now Malaysia were developed.
This was a far superior location for trans-peninsular crossings. In fact, it is arguably the single-best location for engaging in international trade in all of Asia.
The presence of large lakes (Lake Songkhla, Thale Noi, etc.), river systems, and two sea coasts, made this a much more effective location for crossing the peninsula (as well as extracting resources and engaging in trade with people living throughout this extensive water world) and it is therefore not surprising that the largest number of Tang-era ceramics have been found on either side of the Malay Peninsula in this area.
I have written at length about this, and there is no doubt in my mind that we can see Chinese and Arabic sources, among others, all referring to this location (see my working paper for details).
For some reason, as this area became a major trading hub, foreigners came to refer to this place as “Java,” or possibly “Jaba.” This is what the “Shepo” in Chinese referred to. Perhaps that vague older term that we find in some Chinese sources was used in a new way, and then adopted by other foreign merchants? I’m not sure.
Later, in the 1200s and 1300s, in Chinese and Arabic sources, as well as in Marco Polo’s writings, we see references to “two Javas.” This happened, I argue, as trade and contact extended further into the island world towards the Spice Islands (and I have blog posts on this).
Much like the way that foreigners would later use the term “Siam” to refer to a place that the “Siamese” did not refer to by that name, and much like the way that “Siam” was a confederation of city-states with one powerful center, so was “Java/Jaba” a confederation of city states with a powerful center which the locals did not refer to as “Java/Jaba.”
This is clearly the area that the Chola attacked in the early eleventh century. They referred to the powerful center as Srivishajam/Srivijayam, the same name that appears in late-seventh century inscriptions in southern Sumatra and on the 775 Ligor inscription in southern Thailand.
This term, srivijaya/srivishaya, can mean something like “the royal capital.” One possibility is that this is the way that the ruling elite of the powerful center at the heart of the Songkhla-Kedah trading empire referred to their capital. Whether or not they also used this name to refer to something larger than the capital, however, is unknown.
The Chola also referred to a ruling line called the Śailendra, the same name that is mentioned on the Ligor inscription as well as in eighth-century inscriptions (or an inscription?) on the island of Java. I do not have the ability to read the inscriptions and evaluate the evidence for the Śailendra on Java, so I cannot make the argument that this was one ruling family, however, from the research that I have done to date, I would suggest that it is MUCH MORE LOGICAL to argue that there was a “Java/Jaba expansion” from the trans-peninsular trading empire of Java/Jaba in the Lake Songkhla area starting in seventh and eighth centuries, just as that area was starting to prosper, than to argue for an expansion in the other direction. After all, many of the inscriptions are in Old Malay. . .
The attacks on the Southeast Asian mainland and the threat of “Java/Jaba” to Cambodia is also more logical from this perspective. The idea that a new trading empire based at the most geographically prime area in Asia for international trade would seek to expand its control and influence over other polities is also logical, and the idea that foreigners, like the Chola, would be envious and seek to attack this place is likewise logical.
Song dynasty era sources say that Sanfoqi and Shepo were huge rivals. In the current orthodox scholarly view, “Sanfoqi” is a place called “Srivijaya” located at Palembang on Sumatra and “Shepo” is island “Java.”
Through extensive documentation (and there is more on the way), I have demonstrated that this is not accurate. It is totally clear to me that Sanfoqi was Kambuja and Shepo was Java/Jaba, a name that foreigners used to refer to a trade empire with its center in the Songkhla region.
However, let’s put that aside for the moment, and just use our brains and our common sense.
From the current perspective, the rivalry between Sanfoqi and Shepo would mean that a thousand years ago, when Angkor was at its peak (as we can see from inscriptions and archaeology), and when Kedah was a major anchorage for ships coming from the west (as we can see from Arabic sources), a place called Srivijaya at Palembang (for which the scholarship is astoundingly weak and problematic) and island Java were great rivals, and this is something that was considered important to record in Chinese sources (it’s presented as THE big rivalry).
Meanwhile, we can see in Arabic sources, for instance, that their merchants didn’t even visit Palembang or island Java. When they sailed all the way to China, they went from Kedah, to Tioman Island to get water, and then to Qmar (Cambodia), etc. . .
Further, the information about “Zhenla,” the place that scholars think was Cambodia/Angkor, is extremely sparse. As such, if we follow the orthodox view of the past, then the area from the southern coast of Cambodia across to the eastern side of the Malay Peninsula is “empty” in Chinese sources for the period when Angkor flourished.
In other words, there is basically no Angkor/Cambodia in Chinese sources. Instead, the big story is a supposed rivalry between Palembang and island Java. . .
Make that make sense.
Again, I have already provided extensive documentation that demonstrates that this was not the case. I’ve also said many times that when Georges Cœdès claimed that Sanfoqi was a kingdom called “Srivijaya” on the island of Sumatra, it was as though he took a map and hung it on a wall upside-down. Since then, everyone has tried to connect information that they find in historical sources with the places on the map, not realizing that the map is upside-down.
Lowman lays out key historical information in his dissertation, but try as he may, he can’t get that information to lead to a clear conclusion, because, ultimately, he can’t find an alternative to Cœdès’s upside-down map to connect the information to. You can tell that he tries very hard to do so, and he’s very logical, but in the end, he just can’t succeed in figuring out where the Java that Jayavarman II reportedly fled from was located.
I really admire the effort that Lowman made in his dissertation. Wisely, he largely avoided the writings about Chinese sources. There is no O. W. Wolters or Paul Pelliot in this chapter, for instance, and that is good, because the “orthodox” secondary scholarship on Chinese sources is all too flawed to employ, and can only lead readers down a rabbit hole of misunderstandings and bad ideas.
Yes, there was a “Java/Jaba” expansion, and it affected the Khmer. This was not a myth. It came from the center of the Malay Peninsula, and it was triggered by the development of trade routes across the peninsula in the Songkhla/Phatthalung – Kedah/Trang region in the years before 802 when Jayavarman II “came from Javā” (although, what exactly that means is still unclear) and had a ritual performed that “made it impossible for this [country of] Kambujadeśa to be dependent on Java, and made it so that there would be only one king who was cakravartin (‘universal ruler’).”
This was the Shepo – Sanfoqi rivalry that would soon be recorded in Song dynasty era sources. As such, the area from the southern coast of Cambodia across to the eastern side of the Malay Peninsula is not actually “empty” in Chinese sources. It has only been rendered that way by modern scholars who have based their work on Cœdès’s upside-down map.
To the contrary, in Song dynasty era sources, the area of the Gulf of Thailand between Ha Tien, the main trade port for the Cambodian world, and Songkhla and Sathing Pra was the place where the main action was.
What is “empty” in Chinese sources up until the Mongol period is the area around southern Sumatra and island Java. It was only when the Mongols attacked Java, and the subsequent Ming dynasty established a tributary relationship with a kingdom there that island Java (or Zhuawa, as it is rendered) entered the Chinese historical record.
This is very logical. The Chinese knew more about places closer to China before they knew about places farther away from China.
The Khmer were not the only people whom Shepo/Java/Jaba attacked. The late 7th century, Srivijaya inscriptions in southern Sumatra (in Old Malay) are a sign of their southern expansion, as is perhaps the presence of a Śailendra “dynasty” on Java. The area of northern Vietnam was attacked in 767 by people from Shepo (Songkhla area) together with Kunlun 崑崙 people, a reference to the area around the southern tip of the Indochinese Peninsula. Phan Rang was attacked in 787 by people from Java.
Then you have Jayavarman II moving away from some kind of control by Java in 802, and the Chola attacking the area of Java/Jaba at Songkhla in the early eleventh century.
Indeed, there was definitely a lot of action around the Gulf of Thailand and beyond during this time period. However, all of the pieces fit together once we turn Cœdès’s up-side-down map right-side-up by recognizing that Sanfoqi was Kambuja and Shepo was Java/Jaba, a trading center at the heart of the trans-peninsular crossing between Songkhla and Kedah.
Although this is not what Lowman concluded, I still find his dissertation to be fantastic for the way that he clearly and logically laid out all the evidence, and thought through it.
Without knowing how badly scholarship has been distorted by Cœdès’s upside-down map, there was no way for Lowman to make a conclusion about where the “Java/Jaba” that rivaled Kambuja was located.
Now, however, that the map is right-side-up, we can see, and we can also see how all of the historical information fits together and makes sense.
Hi Prof. I will be more agree that the attack from Java upon Champa ini 787a.d. is basically came from Malay Peninsula. Not from Java Island (Indonesia).
Thank you for the comment and for agreeing. What makes you think that way?
Wanna know why I did agree with you Prof? Because we know that when the Sri Maharaja expanded his power as far as Ligor (Nakhon Si Thammarat), if we refer to the findings of the Ligor Inscription, the Sri Maharaja likely relied on military assistance from the Malay Peninsula, particularly in the Kra Isthmus, to launch an attack on Champa.
To me, it would have been impossible at that time to bring thousands of troops directly from the island of Java to the Malay Peninsula before continuing their journey to Champa. Therefore, it seems more logical that they recruited people from around the Javaka region to carry out the attack.